Replace Plaud sync timeouts with inactivity-based approach
Remove the 3600s wall-clock subprocess timeout that killed active,
progressing imports (e.g. 26/28 segments done but >1hr elapsed).
The inner import process already has its own 600s inactivity timeout
for stall detection, making the outer backstop redundant.
Convert the overall sync timeout from wall-clock to inactivity-based:
timer now resets after each file completes (success or failure), so
a batch of long imports will process fully as long as each one
eventually finishes.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>